Please use this identifier to cite or link to this item: http://hdl.handle.net/2080/4389
Title: Deep Learning Approach for Directivity Prediction of Multi-Layered Leaky Wave Antenna
Authors: Mazumder, Lopamudra
Ghosal, Sandip
De, Arijit
Keywords: Leaky wave antenna
Multi-layered medium
Directivity
Deep learning
Issue Date: Nov-2023
Citation: IEEE 3rd International conference on Applied Electromagnetics, Signal Processing & Communication (AESPC), KIIT University, 24-26 November 2023
Abstract: This work provides a solution for the forward design problem of leaky wave antennas to predict its directivity using a novel deep neural network architecture. Considering the nonlinear relation of the angle of maximum directivity and frequency at leaky wave resonance, this paper leverages the advantage of deep learning framework for such non-linear regression problem. Using the Maxwell’s equations, the dataset has been generated for different permittivity and thickness values of three-layered leaky wave antenna. For training purpose, it considers 80% of total data. Remaining 20% data are used to obtain an accurate prediction of the maximum directivity within less than 1% error. The proposed method can be suitable replacement for rigorous full-wave simulation in terms of time and computational resource.
Description: Copyright belongs to proceeding publisher
URI: http://hdl.handle.net/2080/4389
Appears in Collections:Conference Papers

Files in This Item:
File Description SizeFormat 
2023_AESPC_LMazumder_Deep.pdf318.88 kBAdobe PDFView/Open    Request a copy


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.